The Application Of Recurrent Neural Network Model In Nonlinear Neutralization Process Control With Pure Time

Neutralization process is a typical nonlinear process with pure time delay. This process can't be controlled effectively with conventional linear control method. Nonlinear gain compensation control and feed forward control are employed in the practical application. The realization of thir control system and its control result are introduced in this paper. Optimal control can't be realized for this method is not a real nonlinear control method.Neural networks can be viewed as a universal approximator for nonlinear functions, but the multilayer feedforward neural network which be used usually is a static state network in nature, it is disagree with the real-time identification for dynamic system. Moreover, recurrent neural networks can simulate the state memory mechanism of dynamic system, so it can be utilized as the model of dynamic time delay system. In this paper, one step ahead predtictive control strategy based on neural network (NMPC) is presented. Model training algorithm using Elman nets is described in details, also does NMPC algorithm using network's gradient information. In addition. the realization of whole neural network predictive control system and the control results are given. Real time control test show that this control system have good dynamic performance and excellent robustness.At last, a summary of the paper is given and further research interests are also proposed.